Main Report
Summary
Deep repository for spent nuclear fuel
SR 97 – Post-closure safety
Foreword
During the past three years, SKB has carried out an assessment of the long-term safety of a deep repository for spent nuclear fuel. The results of the project are reported in Swedish as “Djupförvar för använt kärnbränsle; SR 97 – Säkerheten efter förslutning”. This report is an English translation titled “Deep repository for spent nuclear fuel; SR 97 – Post-closure safety”. The Main Report in its complete form consists of two parts with accounts of premises, methodology, analyses, results and conclusions. In addition there is a detailed summary which contains, among other things, the entire conclusion chapter from the complete version.
The report is primarily written for experts, but parts of the text should be of interest to non-specialists as well.
Allan Hedin has been responsible for methodology and for coordination of the different parts of the project, has written the summary, and has acted as writing editor for the complete main report. Patrik Sellin has dealt with near-field subjects. Anders Ström and Jan-Olof Selroos have been in charge of geosphere-related matters, and Ulrik Kautsky has been responsible for the biosphere. Lena Morén has worked with the climate and intrusion scenarios, while Fredrik Lindström has carried out the radionuclide transport calculations.
Many other individuals inside and outside SKB have also contributed in various ways to the project. If any are to be given special mention, the difficult choice falls on Johan Andersson of Golder Grundteknik, who participated as an expert in both geosphere matters and safety assessment in general, and Harald Hökmark of Clay Technology, who has worked with mechanical questions in the geosphere.
SKB is responsible for all judgements and conclusions in the report.
Stockholm, November 1999
Tönis Papp
Contents
Summary 7 1 Premises 11 1.1 Why SR 97? 11 1.2 Purposes 12 1.3 Delimitations 13 1.4 Report structure 14 1.5 Acceptance criteria 151.5.1 SSI’s regulations for final disposal of spent nuclear fuel 15 1.5.2 SKI’s draft version of regulations concerning safety in final
disposal of nuclear waste 17
1.6 Safety principles 17
1.7 Time perspective 18
1.7.1 Time periods in SR 97 19
1.8 Methodology 19
1.8.1 What is a safety assessment? 19
1.8.2 Handling of uncertainties 20
2 System description/initial state 23
2.1 Methodology 24
2.1.1 System description of THMC format 24
2.2 Fuel 25
2.3 Cast iron insert/copper canister 27
2.4 Buffer/backfill 28
2.5 Geosphere/site descriptions 29
2.5.1 Crystalline rock 29
2.5.2 The three sites in SR 97 29
2.5.3 Site-adapted repository layouts 33
2.5.4 THMC description of processes and variables 35
2.6 Safety criteria 35
2.7 Completeness of system description 36
3 Scenarios 39
3.1 Choice of scenarios 39
3.1.1 Completeness/coverage in choice of scenarios 41
3.2 Base scenario 42
3.2.1 Initial state and boundary conditions 43 3.2.2 Overview of processes and dependencies 43
3.2.3 Radiation-related evolution 44
3.2.4 Thermal evolution 46
3.2.5 Hydraulic evolution 48
3.2.6 Mechanical evolution 51
3.2.7 Chemical evolution 53
3.2.8 Summary: The base scenario in a time perspective 57
3.3 Canister defect scenario 60
3.3.1 Initial canister defects 61
3.3.2 Radiation-related evolution, criticality 61 3.3.3 Hydromechanical evolution in defective canister 63 3.3.4 Chemical evolution in defective canister 66 3.3.5 Hydraulic evolution in the geosphere 70 3.3.6 Transport processes in the repository 74
3.3.7 Biosphere 75
3.3.8 Calculations of radionuclide transport 78
3.3.9 Discussion of results 88
3.4 Climate scenario 92
3.4.1 Climate scenario in SR 97 93
3.4.2 Impact on the repository 94
3.4.3 Thermal evolution 94 3.4.4 Hydraulic evolution 94 3.4.5 Mechanical evolution 94 3.4.6 Chemical evolution 95 3.4.7 Radionuclide transport 96 3.4.8 Summary 96 3.4.9 Uncertainties 97 3.5 Earthquake scenario 97 3.5.1 Background 97
3.5.2 Analysis of earthquake risks 99
3.5.3 Uncertainties 100
3.5.4 Conclusions 101
3.6 Intrusion scenario 101
4 Discussion and conclusions 103
4.1 Safety of KBS-3 method in Swedish bedrock 103 4.1.1 Are all internal processes and external events of importance
identified? 104
4.1.2 What are the results of the different scenario analyses and
what confidence can be attached to the results? 104 4.1.3 Weighing-together of scenario analyses 108 4.1.4 How do different conditions in Swedish bedrock affect the
feasibility of building a safe repository? 109
4.2 Methodology for safety assessment 112
4.2.1 System description 112
4.2.2 Choice of scenarios 113
4.2.3 Analysis of chosen scenarios 113
4.2.4 Handling of uncertainties 114
4.2.5 Assessment of available methodology 115 4.3 Basis for site selection and site investigations 115
4.3.1 What requirements does the deep repository make on the
host-rock? 115
4.3.2 Programme for site investigations 116
4.4 Basis for functional requirements 116
4.5 Prioritization of research 118
Summary
Purpose and premises
In preparation for coming site investigations for siting of a deep repository for spent nuclear fuel, the Swedish Government and nuclear regulatory authorities have requested an assessment of the repository’s long-term safety with the following purpose: “…to demonstrate that the KBS-3 method has good prospects of being able to meet the safety and radiation protection requirements which SKI and SSI have specified in recent years.” SR 97 is the requested safety assessment. The purpose is to demonstrate by means of a systematically conducted analysis whether the risk of harmful effects in individuals in the vicinity of the repository complies with the acceptance criterion formulated by the Swedish regulatory authorities, i.e. that the risk may not exceed 10–6 per year. Geological
data are taken from three sites in Sweden to shed light on different conditions in Swedish granitic bedrock. The repository is of the KBS-3 type, where the fuel is placed in isolating copper canisters with a high-strength cast iron insert. The canisters are surrounded by bentonite clay in individual deposition holes at a depth of 500 m in granitic bedrock.
The assessment applies to a closed repository for spent nuclear fuel and thus does not include either safety during operation or safety of the repository for long-lived low- and intermediate-level waste. These matters are dealt with in separate reports.
Methodology
The methodology in the assessment entails first describing the appearance of the repository when it has just been closed and then analyze how the system changes with time as a result of both internal processes in the repository and external forces. The future evolution of the repository system is analyzed in the form of five scenarios. The first is a base scenario where the repository is postulated to be built entirely according to specifications and where present-day conditions in the surroundings, including climate, are postulated to persist. The four other scenarios show how the evolution of the repo-sitory differs from that in the base scenario if the reporepo-sitory contains a few initially defective canisters, in the event of climate change, in the event of earthquakes, and in the event of future inadvertent human intrusion. Repository evolution is broken down into thermal, hydraulic, mechanical and chemical processes, and the ultimate purpose of the analyses is to evaluate the repository’s capacity to isolate the waste in the canisters, and to retard any releases of radionuclides if canisters are damaged. The time horizon for the analyses is at most one million years, in accordance with preliminary regulations.
Base scenario
By means of model studies and calculations, the base scenario analyzes how the radio-toxicity of the fuel declines with time, the repository’s thermal evolution as a result of the decay heat in the fuel, the hydraulic evolution in buffer and backfill when they become saturated with water, and the long-term groundwater flow in the geosphere on the three sites.
Mechanical stresses on the canister stemming from groundwater pressure and swelling pressure from the buffer are examined, along with the long-term mechanical stability of the geosphere. The chemical evolution of bedrock and buffer, as well as corrosion of the copper canister, are also analyzed.
The overall conclusion of the analyses in the base scenario is that the copper canister’s isolating capacity is not threatened by either the mechanical or chemical stresses to which it is subjected. The safety margins are great even in a million-year perspective.
Canister defect scenario
The internal evolution in initially defective canisters and the possible resultant migration of radionuclides in buffer, geosphere and biosphere are analyzed in the canister defect scenario. The result is estimates of dose and risk that can be compared with the accept-ance criterion for a deep repository.
The scenario first shows that criticality cannot be expected to occur in the repository. Analyses of the hydromechanical evolution in a damaged canister when water enters show that even the damaged canister prevents the release of radionuclides for a very long time, since intruding water is consumed by corrosion of the cast iron insert. Dissolution of the fuel and solubility conditions for radionuclides released from the fuel are studied in analyses of the chemical evolution in a damaged canister. Model calculations show that hydrogen gas generated by corrosion of the cast iron insert contributes towards keeping the rate of fuel dissolution low.
Groundwater flow is studied on a local scale on the three sites. The analyses show that variation in results stemming from the natural variability in the rock often overshadows the variation caused by both differences between model concepts and uncertainties in boundary conditions, fracture structure, etc.
Radionuclide flux in the biosphere is modelled for a number of ecosystems, e.g. well and peatland. Peatland gives relatively high doses as a consequence of accumulation of e.g. Ra-226.
Data from the above-mentioned studies are then used for calculations of radionuclide transport in canister, buffer, backfill and geosphere. Releases from the geosphere are converted to doses in different ecosystems. Both reasonable and pessimistic values are estimated for all input data to the calculations, and in a few cases statistical distributions as well.
With reasonable data, the doses on all sites lie far below the dose limits that can be derived from the official acceptance criteria. The influence of uncertainties in data is analyzed by systematically substituting reasonable data for pessimistic data and studying the calculation result. The variation in flow-related data in the geosphere has the greatest impact on the result, followed by data uncertainties for the biosphere. Other conclusions are that our understanding of fuel dissolution needs to be improved, and that the
probability and size of initial canister defects that escape quality-control inspection is difficult to estimate.
In order to obtain a risk measure that can be directly compared with the acceptance criterion, risk analyses in the form of simplified probabilistic calculations are also performed. The risk analyses show that all sites lie well below the acceptance criterion.
The maximum risk for release to a well is never more than 0.5 percent of the acceptance criterion, even when the calculations are extended a million years into the future. The same applies to releases to peatland for times up to 100,000 years, while the maximum risk here grows to about one-tenth of the acceptance criterion at the least favourable site at times after 100,000 years.
Climate scenario
The consequences of future climate change are explored in the climate scenario. Today’s climate is relatively warm by historical standards, and future changes are expected for the most part to lead to a colder climate as a consequence of cyclical variations in insolation. A conceivable sequence of events, including severe glaciation, on each of the three sites is sketched for the coming 150,000 years.
The repository system’s thermal, hydraulic, mechanical and chemical evolution under the changed conditions in the surroundings is studied in the form of a comparison with the evolution in the base scenario.
In the climate scenario as well, the overall conclusion is that the isolating capacity of the copper canister is not threatened by either mechanical or chemical stresses. The mechanical stresses are larger than in the base scenario, mainly due to higher rock and groundwater pressures in connection with a glaciation. The chemical stresses are roughly the same, partly because oxygen-containing groundwater is not expected to reach the canister. The strength calculations for the canister may need to be refined with more realistic, inhomogeneous material properties, and buffer erosion with extremely ion-poor groundwater compositions may require further study.
As far as the retarding capacity of the repository is concerned, for example in the event of initial canister damage, the most important changes take place in the biosphere. The repository sites are expected to be covered by ice sheets or sea during long periods, and the aggregate effect of climate change will therefore be a reduction of the dose
consequences compared with a situation where the present-day climate persists.
Earthquake scenario
In the earthquake scenario, the consequences of earthquakes are analyzed by means of model studies where site-specific data are used for the structure of the geosphere and for earthquake statistics. The analysis method is new and includes several highly pessimistic simplifications. The analyses show that the probability of canister damage is comparable with the probability assumed for initial damage in the canister defect scenario. In the evaluation of the analysis method, it is shown how less pessimistic assumptions should lead to no canister damage at all in the model studies. The method will be refined.
Intrusion scenario
The scenario that deals with future inadvertent human actions that could conceivably affect the repository is surrounded by great uncertainties, chiefly because the evolution of human society is in principle unpredictable. SR 97 discusses how conceivable societal evolutions and future human actions that affect the repository can nevertheless be categorized to some extent. In an illustrative example, a situation is analyzed where a canister in the repository is inadvertently penetrated by rock drillers. Dose and risk are
point in time. The risk to both drilling personnel and family is judged to lie well below the acceptance criterion, since the probability of the analyzed events is estimated to be very small.
Conclusions
The principal conclusion of the SR 97 safety assessment is that the prospects of building a safety deep repository for spent nuclear fuel in Swedish granitic bedrock are very good. The three analyzed sites reflect reasonable variations of the conditions in granitic
bedrock in Sweden. The analysis does not provide support for attaching any significant importance to differences in long-term safety between sites in a weighing together of all the factors that influence the siting of a deep repository.
Another conclusion is that the methodology that is used in SR 97 comprises a good foundation for future safety assessments that will be based on data from completed site investigations.
The results of the assessment also serve as a basis for formulating requirements and preferences regarding the bedrock in site investigations, for designing a programme for site investigations, for formulating functional requirements on the repository’s barriers, and for prioritization of research.
The next stage in the siting of a deep repository entails investigation of the bedrock at a number of candidate sites in Sweden. It is SKB’s judgement that the scope of the safety assessment and confidence in its results satisfy the requirements that should be made in preparation for such a stage.
1
Premises
Under Swedish law, the owners of nuclear reactors are obligated to see to it that radioactive waste from their activities is managed and disposed of safely. The Swedish power utilities jointly own Svensk Kärnbränslehantering AB, SKB (the Swedish Nuclear Fuel and Waste Management Company), whose mission is to develop methods for managing radioactive waste and to build and operate the facilities required for this. Spent nuclear fuel is an important component in the radioactive waste, since it is both highly radioactive (high-level) and long-lived. At present, spent fuel is stored for a year or so at the reactor, after which it is transferred to CLAB, a central interim storage facility for spent nuclear fuel. According to SKB’s plans, after 30 to 40 years of interim storage the fuel will be encapsulated in copper canisters and disposed of at a depth of approximately 500 metres in the crystalline bedrock. The facilities required for this, an encapsulation plant and a deep repository, have not yet been sited and built.
The system will be constructed over a period of several decades. Siting of facilities and systems is done in collaboration with concerned municipalities and under the super-vision of safety and radiation protection authorities, all subject to the approval of the Government.
1.1
Why SR 97?
In preparation for the next stages in the realization of the system, the Swedish Government stated the following in its decision following the review of SKB’s research programme RD&D 95:
“A safety assessment of the repository’s long-term safety should, in the opinion of the Government, be completed before an application for a permit to construct an encapsulation plant is submitted, likewise before site investigations on two or more sites are commenced.”
This report gives an account of the requested safety assessment before site investigations are commenced. The working title of the analysis is SR 97 (Safety Report 97).
In its review of SKB’s RD&D 98, the Swedish Nuclear Power Inspectorate (SKI) clarifies its view of the purpose and requirements for SR 97:
“The purpose is to demonstrate that the KBS-3 method has good prospects of being able to meet the safety and radiation protection requirements which SKI and SSI have specified in recent years.”
SKI also writes: “…that SR 97, besides demonstrating a methodology for safety assessment, should also serve as a basis for:
• demonstrating the feasibility of finding a site in Swedish bedrock which meets the requirements on long-term safety and radiation protection that are defined in SSI’s
• specifying the factors that serve as a basis for the selection of areas for site investigations,
• deriving which parameters need to be determined and which other requirements ought to be made on a site investigation,
• deriving preliminary functional requirements on the canister and the other barriers.”
1.2
Purposes
Based on the above points, four concrete purposes for SR 97 can be formulated: 1. SR 97 shall serve as a basis for demonstrating the feasibility of finding a site in
Swedish bedrock where the KBS-3 method for deep disposal of spent nuclear fuel meets the requirements on long-term safety and radiation protection that are defined in SSI’s and SKI’s regulations.
2. SR 97 shall demonstrate methodology for safety assessment.
The ambition of SR 97 is to carry out a complete analysis of the long-term safety of the KBS-3 system for deep disposal of spent nuclear fuel. The methodology employed in SR 97 includes:
• a systematic handling of all the internal processes and external conditions that can cause long-term changes in the repository, and
• a systematic handling of the different types of uncertainties that always surround the background data for an analysis.
SR 97 is based on data from three actual sites. Data have been taken from SKB’s investigations at Gideå in Ångermanland, from Finnsjön in northern Uppland County and from the Hard Rock Laboratory on Äspö outside Oskarshamn in Småland. The sites have been selected as calculations examples to reflect different conditions in Swedish granitic bedrock as regards geology, groundwater flux, water chemistry, nearness to coast, northerly or southerly location, surrounding biosphere, etc.
The report on the execution and results of the analysis therefore serves as a direct basis for assessing: a) the feasibility of finding a safe site for a KBS-3 repository in Swedish bedrock, and b) the methodology for a safety assessment.
3. SR 97 shall serve as a basis for specifying the factors that serve as a basis for the selection of areas for site investigations and deriving which parameters need to be determined and which other requirements ought to be made on a site investigation. SR 97 comprises an important supporting document in the ongoing work of formulating requirements and preferences regarding the rock from the perspective of longterm safety. Results and experience from SR 97 are also used directly in the work of formulating an integrated programme for investigations and evaluations of sites. The conclusion chapter summarizes the way in which SR 97 comprises a background document for these two efforts.
4. SR 97 shall serve as a basis for deriving preliminary functional requirements on the canister and the other barriers.
How functional requirements can be derived from the results of the safety assessment is discussed in the conclusion chapter.
1.3
Delimitations
SR 97 is a complete safety assessment of the KBS-3 method for deep disposal of spent nuclear fuel, where geosphere data are taken from three actual sites in Sweden. The following fundamental premises also apply:
Post-closure safety
SR 97 deals with the long-term safety of the repository after closure. The construction and operating phases are not dealt with. These phases, as well as other aspects that pertain to the whole waste management system (encapsulation, transportation and deep disposal), are described in preliminary safety reports in conjunction with construction and operation. Together with SR 97, they comprise the background material for an integrated system analysis of all components in the waste management system to be published in 2000. Nor is SR 97 concerned with safety in connection with a prolonged open period or a partially closed repository.
Repository for spent nuclear fuel
SR 97 is concerned with a repository for spent nuclear fuel. Other long-lived waste will also have to be disposed of, for example core components from the decommissioning of nuclear power plants and waste from previous activities at the research reactor at Studsvik. This waste will be emplaced in a separate repository, which can be co-sited with the repository for spent nuclear fuel or with the final repository for radioactive operational waste, SFR, which is in operation today. The repository can also be sited separately.
A preliminary facility design and safety assessment for such a repository has been prepared in parallel with SR 97 and is presented in a separate report /SKB, 1999/. The safety-related consequences of a possible co-siting are not investigated in SR 97, but both assessments are based on the same geological data.
Holistic view of radiation protection
Long-term post-closure safety is one aspect of a holistic view of radiation protection in connection with waste management. A complete picture is presented in the system analysis mentioned above. The options allowed within the frames for KBS-3, as well as when and based on what information they will be evaluated and screened, are also discussed in that report.
1.4
Report structure
The structure of the account in SR 97 represents a development of the template devised in 1995 in safety report SR 95.
The body of material for a safety assessment is very large. SR 97 is presented in the form of a main report to which three main references are closely associated, see Figure 1-1. In the main report and the three main references, reference is made to reports in SKB’s report series or in the open literature.
The main report – ”Deep Repository for Spent Nuclear Fuel; SR 97 – Post-closure safety” – summarizes the entire safety assessment. It can be read separately from the others and includes methodology description, all essential results, as well as evaluations and conclusions. The report consists of two parts and a summary (this volume). All parts are available in Swedish and English.
“SR 97 – Waste, repository design and sites” describes in detail the waste, the repository design with canisters and buffer/backfill material, the three sites and the site-specific adaptations of the repository layouts that have been done. The report is available in both Swedish and English. Hereinafter, this report will be referred to as the “Repository System Report”.
“SR 97 – Processes in the repository evolution” describes the thermal, hydraulic, mechanical and chemical processes in fuel, canister, buffer and geosphere that control the evolution of the repository system. The report is available in both Swedish and English. Hereinafter, this report will be referred to as the “Process Report”. “SR 97 – Data and data uncertainties” (in English only) contains a compilation of input data for calculations of radionuclide transport. There is also an evaluation of uncertainties in input data. Hereinafter, this report will be referred to as the “Data Report”.
1.5
Acceptance criteria
The form and content of a safety assessment, and above all the criteria for judging the safety of the repository, are defined in regulations issued by the Swedish safety and radiation protection authorities. The regulations are based on framework legislation, the most important being the Environmental Code, the Nuclear Activities Act and the Radiation Protection Act. Radiation protection matters are handled by a number of international bodies, and national legislation is often based on international rules and recommendations.
Long-term safety is regulated today by the Swedish Radiation Protection Institute’s (SSI) “Regulations for final disposal of spent nuclear fuel” (SSI FS 1998:1). The regulations entered into force on 1 February, 1999.
In 1999, the Swedish Nuclear Power Inspectorate, SKI, distributed a draft version of “The Swedish Nuclear Power Inspectorate’s regulations concerning safety in final disposal of nuclear waste”.
1.5.1
SSI’s regulations for final disposal of spent nuclear fuel
SSI writes that human health and the environment, now and in the future, shall be protected from the harmful effects of ionizing radiation. Nuclear activities must not cause more serious effects on human health and the environment outside Sweden’s boundaries that what is acceptable within Sweden. A final repository shall be designed so that no additional measures are needed after closure to prevent or limit the escape of radioactive substances from the repository. Institutional controls and knowledge of the location of the repository in a distant future cannot be assumed. SSI’s regulations apply to the long-term safety of a closed repository.
Protection of human health
The overall acceptance criterion for a deep repository is expressed in Section 5 of SSI’s regulations:
“A final repository for spent nuclear fuel or nuclear waste shall be designed so that the annual risk of harmful effects after closure is no more than 10–6 for a representative
individual in a group that is exposed to the greatest risk.”
The acceptance criterion is thus a risk measure. A risk calculation investigates what courses of events can lead to harmful effects, what their probability of occurring is, and the size of the injury (the consequence) for each course of events. The product of probability and consequence gives a sub-risk for each course of events. The aggregate risk is the sum of the sub-risks for different conceivable courses of events.
SSI stipulates an annual risk of 10–6 for individuals exposed to radiation from the
repository. For a hypothetical situation with exposure that occurs with certainty (probability = 1), this corresponds to an annual radiation dose of 0.015 milliSieverts (mSv) from the repository. This can be compared with the natural background radiation, which is several mSv/y in Sweden.
The risk limit applies to a representative individual in the group that is exposed to the greatest risk. As an indication of the size of such a group, SSI mentions the population in an area where it is theoretically possible to site ten different deep repositories. Such an
acceptable to carry out the calculations for an individual judged to be highly burdened, instead of an individual who is representative for the whole group’s burden”.
The risk limit for such an individual is set at 10–5, which corresponds to a radiation dose
of 0.15 mSv/y. The exposure models in SR 97 have not been adapted to the details of SSI’s regulations, since the latter did not enter into force until towards the end of the assessment. However, the models are already designed in most cases to calculate doses to a small and highly exposed group which, for example, lives solely on contaminated food. The calculation result in SR 97 should therefore in most cases be compared with the risk criterion 10–5/y, equivalent to a dose limit of 0.15 mSv/y for an exposure that occurs with
certainty, see further section 3.3.7.
Environmental protection SSI also states that:
“§7 An account shall be given of biological effects of ionizing radiation in affected habitats and ecosystems. The account shall be based on available knowledge of concerned ecosystems …”
In the absence of established methodology, SSI says that the precautionary principle shall apply, i.e. the very suspicion of harmful effects on the environment shall be sufficient to intervene or refrain from a given activity.
In SR 97, the biological effects of a release are judged by comparison with the natural background radiation. If the releases are small compared with the background radiation, the effects should be negligible.
Intrusion
SSI stipulates that an account shall be given of the consequences of an inadvertent intrusion or other disturbance in the final repository or its vicinity. What is essential is not to describe the chain of events leading up to the intrusion, but to shed light on the repository’s protective function after an intrusion. The protective capacity of a final repository must not be impaired by planned measures to hinder intrusion or facilitate retrievability.
Doses higher than 1 mSv/y, which could conceivably be encountered in connection with an intrusion into the final repository, will be assessed separately by SSI.
Time periods
SSI states that harmful effects in the future should not be regarded as less important than the harmful effects to which man or the environment are exposed today.
SSI emphasizes that the first 1,000 years after repository closure is the most important period to investigate, since the radiotoxicity of the waste is greatest then. The highest demands are made on the safety account for this period. The regulations also require an account of a case based on the assumption that the biosphere conditions prevailing at the time of the licence application do not change. The term “prevailing conditions” also takes into account known changes such as postglacial land uplift.
The period after the initial 1,000 years shall also be investigated, and SSI emphasizes the importance of accounting for the different types of uncertainties in the underlying data on which the analyses of different epochs are based.
1.5.2
SKI’s draft version of regulations concerning safety in final
disposal of nuclear waste
The regulations from SKI are as yet only available in a draft version. Among other things, the regulations talk about how the safety assessment should deal with various internal and external conditions that may have a bearing on safety. SKI emphasizes the importance of a systematic handling of uncertainties, and that the models and data used should be demonstrated to be applicable as far as possible. The assessment must cover the first million years after repository closure.
Since the regulations are not yet available in a final version, it has not been possible to use them as a direct basis for SR 97. In general, it can nonetheless be said that all aspects dealt with in the draft version are also covered in one way or another in SR 97.
1.6
Safety principles
As the work of developing a safe deep repository in Sweden has proceeded, a philosophy has emerged regarding how the radioactive waste in Sweden is to be managed. In brief, it entails the following:
• Long-term safety shall not require future monitoring and maintenance.
• The repository shall be designed to permit possible future measures to modify the repository or retrieve the waste.
• The long-term safety of the repository shall be based on multiple engineered and natural barriers which contribute via different functions to the repository’s total safety. The practical application of this philosophy has resulted in a repository design with a multiple barrier system, the KBS-3 system.
The KBS-3 repository for spent nuclear fuel is designed primarily to isolate the waste. If the isolation function should for any reason fail in any respect, a secondary purpose of the repository is to retard the release of radionuclides. This safety is achieved with a system of barriers that support and complement each other. The safety of the repository must be adequate even if one barrier should be defective or fail to perform as intended. This is the essence of the multiple barrier principle.
Another principle is to make the repository “nature-like”, i.e. to use natural materials for the engineered barriers. Choosing materials from nature makes it possible to judge and evaluate the materials’ long-term stability and behaviour in a deep repository based on knowledge of natural deposits. For the same reason, the repository should cause as little disturbance of the natural conditions in the rock as possible. Above all, an attempt is made to limit the chemical impact of the repository in the rock.
1.7
Time perspective
The repository should function as long as the waste is hazardous. It takes many billions of years before all radioactive material has decayed to stable elements. By then, however, their radiotoxicity has long since declined to levels comparable to the radiotoxicity of the uranium ore originally mined to produce the fuel.
Approximately eight tonnes of natural uranium are enriched to fabricate one tonne of fuel for a Swedish reactor. During reactor operations, the radiotoxicity of the fuel increases as new radioactive substances are formed when uranium nuclei undergo fission. Figure 1-2 shows how the radiotoxicity of the spent fuel subsequently declines with time. After approximately 100,000 years, the radiotoxicity of a tonne of spent fuel is on a par with that of the eight tonnes of natural uranium used in fabricating the fuel.
The figure 100,000 years can therefore be used as a guideline for how long the reposi-tory has to “function”. However, this figure is not an absolute time limit in the evalu-ation of the repository’s safety:
• On the one hand, radiotoxicity declines steadily and has e.g. after a thousand years fallen to about one-tenth of the level at deposition. This is important in the evalu-ation of the repository’s safety: Uncertainty regarding conditions in and around the repository grows with time, but at the same time the radiotoxicity of the fuel diminishes.
• On the other hand, even after 100,000 years there are both small quantities of radio-nuclides that can move relatively easily through the repository’s barriers if the copper canister should be damaged, and larger quantities of low-mobility nuclides.
The safety of the repository thus needs to be evaluated far into the future and constantly in the light of how radiotoxicity declines with time.
Figure 1-2. Toxicity of the waste as a function of time after discharge from the reactor for
Swedish BWR fuel with a burnup of 38 MWd/t U. Radiotoxicity pertains to ingestion via food. After 30 to 40 years of interim storage, the fuel will be deposited in the final repository.
1.7.1
Time periods in SR 97
SR 97 deals with a closed repository. In practice, the repository will be filled in stages over a period of approximately 50 years, but in the analysis it is assumed that all fuel is deposited at the same time. Prior to disposal, it is assumed that the fuel has been kept in interim storage for 40 years after discharge from the reactor.
It is assumed in SR 97 that there are institutional controls for the first one hundred years, which means that inadvertent intrusion can be ruled out during this period.
In accordance with SKI’s above proposed regulations, one million years is the upper time limit for the analyses in SR 97.
1.8
Methodology
1.8.1
What is a safety assessment?
Safety assessment is the method that is used to analyze and judge the performance and safety of a final repository in a systematic manner. A safety assessment of a deep reposi-tory can in simple terms be said to consist of the following tasks:
• carefully describe the appearance or state of the repository system when it has just been closed,
• survey what changes the repository could conceivably undergo in time as a conse-quence of both internal processes within the repository and external forces, • evaluate the consequences of the changes for safety.
In both the execution and the presentation of SR 97, the processes are in focus. Know-ledge of all known internal processes of importance for long-term safety is documented in a special report, the Process Report. The processes and their couplings to each other are illustrated schematically in THMC diagrams, where the processes are classified into the categories thermal (T), hydraulic (H), mechanical (M) and chemical (C). In the analysis, groups of coupled processes are linked together into a description of an inte-grated evolution in time.
The execution and presentation of SR 97 can be divided into five steps: 1. System description
A systematic analysis requires a structured description of the repository and of all the internal processes, their interrelationships and the properties of the repository that are influenced by a particular process. Preparing such a system description is therefore the first task in a safety assessment. This task also includes defining the boundary between a system and its surroundings. The THMC structure is used for the system description in SR 97. The methodology is described more fully in section 2.1.
2. Description of initial state
The initial state of the repository, i.e. what it looks like when it has just been closed, is then described. This includes a description of the dimensions and materials in the engineered barriers as well as the properties of the bedrock around the repository as they appear initially. Gaps in our knowledge of the initial state are also described. 3. Choice of scenarios
The evolution of the repository is influenced by its surroundings. Assessments of the evolution of the surroundings necessarily contain uncertainties: What climatic conditions can be expected in the future? What frequencies and magnitudes of earthquakes can be expected in the repository’s surroundings in the future? To cover different situations in the surroundings, the evolution of the repository is analyzed for a number of different sequences of events in the surroundings: a number of different scenarios are selected and analyzed. The chosen scenarios should together provide reasonable coverage of the different evolutionary pathways the repository and its surroundings could conceivably take.
Scenarios can also be based on different assumptions regarding the initial state in the repository of importance for its long-term evolution.
4. Analysis of chosen scenarios
With the aid of the system description, the evolution of the repository is analyzed for each scenario. A number of different tools and methods are used here, ranging from reasoning and simple approximations to detailed modelling based on site-specific data. In SR 97, a base scenario is first analyzed where the repository is postulated to be built according to specifications and where present-day conditions in the surroundings, including climate, are assumed to persist.
A number of other scenarios are then analyzed where the course of events is compared with that in the base scenario. How will the evolution of the repository change if the climate changes? In the event of earthquakes? If a barrier has a fabrication defect? What importance do these changes have for safety? A basis for carrying out the analyses and reporting them in the form of comparisons with a base scenario is that the analyzed repository system is engineered to be robust, i.e. so that varying conditions in the surroundings do not cause dramatic changes in the evolution and performance of the repository.
5. Evaluation
Finally, an overall assessment is made of repository safety, where the different scenarios are weighed together into a total risk picture. The conclusions of the overall assessment comprise the results of the safety assessment. Confidence in the results in the light of the uncertainties that exist in the data underlying the assessment is also discussed here.
1.8.2
Handling of uncertainties
Just as important as the assessment of the repository’s protective capacity is confidence in the results. The background data for a safety assessment is always burdened with different types of deficiencies. It is, for example, never possible to know in detail the
fracture structure in the host rock, or to achieve certainty about the future climate. Repository safety must be evaluated in the light of such deficiencies. To put it simply, we are faced with the task of showing whether the repository has been designed with adequate margins to be safe in spite of inadequate knowledge. Confidence in the results is dependent on, among other things, how methodically uncertainties and deficiencies have been handled.
Handling of uncertainties is not a separate activity, but comprises an integral part of the analysis work. Uncertainty handling is nevertheless discussed independently here, since it comprises an important part of the methodology.
Deficiencies can be of a qualitative or quantitative nature. Qualitative deficiencies concern e.g. questions of completeness: Have all processes that influence the evolution of the repository been identified in the system description? Have all types of external impact been covered in the choice of scenarios? Other qualitative questions concern process understanding: Do we understand the internal processes well enough for the needs of the safety assessment? Do we understand the processes that determine condi-tions in the surroundings well enough?
Other questions are quantitative. How well can the initial state be determined? The initial temperature of the repository can be determined with an accuracy which is fully adequate for the needs of the analysis, while the description of fracture geometry in the geosphere is burdened with uncertainties that require more careful handling. How well can be describe different processes quantitatively, for example heat conduction or groundwater flow? This question is particularly important for the analysis of radio-nuclide transport, which is of direct importance to the evaluation of the repository’s safety. Calculations of radionuclide transport handle large amounts of input data, which may be burdened with varying degrees of uncertainty.
Handling of uncertainties consists of both reporting uncertainties and deficiencies in the underlying data and handling them when conducting the analysis. Table 1-1 shows the different types of uncertainties which enter into the first four steps of the analysis. As is evident from the table, what is often termed “conceptual uncertainty”, with varying implications in various contexts, has in SR 97 been divided into the concepts “fundamental process understanding” and “model uncertainty”. The former refers to the scientific understanding of a process, while the latter refers to the uncertainties which arise when a process is described with the aid of a model in a safety assessment. Our fundamental understanding of the process of water flow is good, for example. Mathematical models of groundwater movements in fractured rock can nevertheless be formulated in different ways, which entails an uncertainty. The different formulations represent different ways of handling the fact that the details in the nature of the fracture system are only known in a statistical sense.
Uncertainties in data for radionuclide transport
The results of the radionuclide transport calculations quantify repository safety. It is thereby particularly important to assure the quality of these calculations. The handling of uncertainties in input data to radionuclide transport calculations must therefore be rigorous.
Data for transport calculations in SR 97 are taken from background reports by experts within the relevant field. Many of the reports have been specially produced for SR 97.
discussion of uncertainties in the individual reports has then been evaluated in the Data Report. This document has also proposed for all input data:
• a reasonable value and • a pessimistic value.
In addition to a reasonable and a pessimistic value, statistical distributions have also been presented for some input data, such as data that exhibit spatial variability or data determined by spatially variable conditions. In the geosphere, the heterogeneity in e.g. the fracture system gives rise to a distribution in space of various data for calculations, often after extensive analyses.
The span between reasonable and pessimistic values offers an option for representing uncertainties in input data. Radionuclide transport calculations are carried out with different combinations of reasonable and pessimistic data in order to shed light on the importance of different uncertainties.
Furthermore, probabilistic analyses of some kind are required to arrive at a risk measure that can be compared with the acceptance criterion for the deep repository. The follow-ing approach is used in SR 97 for the probabilistic analyses:
• Statistical distributions are used only where there is some kind of statistical material on which to base a distribution, see the examples above.
• Reasonable and pessimistic values are used for other data. Both are assigned probabilities that are chosen so that the risk is deemed to be overestimated in the probabilistic analysis.
Table 1-1. The types of uncertainties that enter into the first four steps of the analysis and where these are presented and handled in the main report summary.
1. System description
Completeness Discussed in section 2.7.
Process understanding Described in detail in a separate report, the Process Report.
(conceptual uncertainty)
2. Description of initial state
Data uncertainty Quantitative uncertainties in the initial state are described for each variable in the complete main report, Chapter 6.
3. Choice of scenarios
Completeness/Coverage Discussed in section 3.1.1.
4. Analysis of chosen scenarios
Model uncertainty General confidence in models used is discussed briefly where model
(conceptual uncertainty) studies are reported in Chapter 3.
Conficence in models for radionuclide transport and groundwater flow is discussed more thoroughly in the complete main report, section 9.11.
Input data to models/ General data uncertainties are discussed briefly where calculations
calculations and model studies are reported in Chapter 3.
Data uncertainties that concern radionuclide transport and groundwater flow are reported and discussed in detail in a special Data Report. Uncertainty analyses and probabilistic analyses are done in the calculations of radionuclide transport.
2
System description/initial state
The KBS-3 repository for spent nuclear fuel is composed of a system of barriers, see Figure 2-1:
• The fuel is placed in corrosion-resistant copper canisters. Inside the five-metre-long canisters is a cast iron insert that provides the necessary mechanical strength. • The canisters are surrounded by a layer of bentonite clay that protects the canister
mechanically in the event of small rock movements and prevents groundwater and corrosive substances from reaching the canister. The clay also effectively adsorbs radionuclides that are released if the canisters should be damaged.
• The canisters with surrounding bentonite clay are emplaced at a depth of about 500 metres in the crystalline bedrock, where mechanical and chemical conditions are extremely stable.
• If any canister should be damaged, the chemical properties of the fuel and the radio-active materials, for example their poor solubility in water, put severe limitations on the transport of radionuclides from the repository to the ground surface. This is particularly true of those elements with the highest long-term radiotoxicity, such as americium and plutonium.
The safety assessment is based on the repository system as it appears at an initial point in time, which in SR 97 is at closure. The way in which the repository system is changed by internal processes and external forces is then analyzed.
A complete analysis requires a formalized description of the repository system and of the processes, with all their couplings and relationships, that control its evolution. The methodology used in SR 97 entails first describing processes, variables and couplings schematically in a system description, after which the variables are assigned initial values, the formalized description of the initial state.
This methodology is applied strictly in the complete version of the main report. In this summary, a simplified account of the methodology for system description is first provid-ed. After that, the initial state of the different barriers is described in greater detail, but in a manner not as strictly formalized as in the complete main report. The description of the geological barrier, the host rock, is at the same time site descriptions of the three sites analyzed in SR 97.
At the end of the chapter, a number of technical criteria are set up against which long-term safety can be evaluated, followed by a discussion of the completeness of the system description.
2.1
Methodology
2.1.1
System description of THMC format
A systematic analysis requires a description of all known internal processes of any con-ceivable importance, their interrelationships and the properties of the repository that are influenced by the particular process. The structure of the description should provide both an overview and details. Another requirement on the structure is that it must be able to be used throughout in the presentation of the safety assessment.
For the system description, the repository is divided into the four subsystems fuel, canister, buffer/backfill and geosphere. All known thermal, hydraulic, mechanical and chemical processes that are of importance for the evolution of the repository are identi-fied for each subsystem. Influences between subsystems, such as buffer and geosphere, are also charted. These influences are also primarily thermal, hydraulic, mechanical or chemical by nature. This method of structuring processes and interactions in the safety assessment is new for SR 97, while the work of identifying the processes included in the structure has been going on for some time.
Variables determine state
The state in a subsystem is characterized at any given moment by a set of variables. The state of the geosphere is, for example, characterized by its temperature, which varies in time and space, by its fracture geometry (which varies widely in space, but hardly at all in time), by groundwater flow, groundwater composition, rock stresses, etc.
Together, the variables should characterize the system sufficiently well to enable a safety assessment to be conducted. Some variables, such as temperature and groundwater composition, are used or determined directly in analyses and calculations, while others serve as a basis for deriving important properties of the system: The thermal conductivity and density of the geosphere can, for example, theoretically be calculated from the variable matrix minerals.
All variables are time-dependent and are influenced by one or more processes, and all processes are influenced by one or more variables.
THMC diagram
All the processes and variables for each subsystem and their interdependencies are gathered into a diagram, which also includes interactions with adjacent subsystems. The diagram is called a THMC diagram, after the classification of the processes and inter-actions into thermal (T), hydraulic (H), mechanical (M) and chemical (C) categories. The THMC diagram for buffer/backfill is shown in Figure 2-2. The diagram also contains radiation-related processes, which have to do with radioactive decay and radiation attenuation in the repository system and processes related to the transport of radionuclides.
Documentation of processes
The system description also includes documentation of knowledge concerning each process. The process documentation constitutes a cornerstone in the background material for a safety assessment. Knowledge of all identified processes in the system description is documented for SR 97 in the Process Report. The following are given for each process: • general description of the process,
• documentation of model studies/experimental studies,
• discussion of uncertainties in both understanding and data for the process,
• proposals as to how the process can be handled for different scenarios in the safety assessment.
2.2
Fuel
The total quantity of fuel obtained from the 12 Swedish nuclear reactors will depend on operating time, energy output and fuel burnup. As of the beginning of 1998, approxi-mately 4,000 tonnes of spent fuel had been generated. With an operating time of 40 years for all reactors, the total quantity of spent fuel can be estimated at 9,500 tonnes. The equivalent quantity for 25 years’ operating time is 6,500 tonnes.
In SR 97 it is assumed that approximately 8,000 tonnes of fuel will be disposed of. It is assumed for the sake of simplification in most subanalyses that all canisters contain fuel from boiling water reactors, BWR fuel, of type SVEA 96 with a burnup of 38 MWd/tU.
Structure of the fuel assemblies
Nuclear fuel consists of cylindrical pellets of uranium dioxide. The pellets are 11 mm high and have a diameter of 8 mm. In SVEA 96 fuel, the pellets are stacked in approxi-mately 4-metre-long cladding tubes of Zircaloy, a durable zirconium alloy. The tubes, or “cans”, are sealed with welds and bundled together into fuel assemblies. Each assembly contains 96 cladding tubes. A fuel assembly also contains components of the nickel alloys Inconel and Incoloy, and parts of stainless steel.
Radionuclides
Radionuclides are formed during reactor operation by nuclear fission of uranium-235 and plutonium-239 in particular, and by neutron capture by nuclei in the metal parts of the fuel. The former are called fission products, the latter activation products. More-over, uranium can form plutonium and other heavier elements by absorbing one or more neutrons. Most of the radionuclides lie embedded in the fuel matrix of uranium dioxide. A few fission products are relatively mobile in the fuel and may migrate to the surface of the fuel pellets during operation.
THMC description of processes and variables
In the complete THMC description of the fuel, important variables are e.g. quantity and distribution of radionuclides, as well as dimensions and material composition of above all the fuel matrix and the cladding tubes. Important fuel-related processes are fuel disso-lution and dissodisso-lution/precipitation of radionuclides.
2.3
Cast iron insert/copper canister
The canister consists of an inner container of cast iron and a shell of copper. The cast iron insert provides mechanical stability and the copper shell protects against corrosion in the repository environment. The copper shell is 5 cm thick and the canister takes the form of an approximately 4.8 metre tall cylinder with a diameter of 1.05 metres.
The insert has channels where the fuel assemblies are placed and is available in two versions: one for 12 BWR assemblies and one for 4 PWR assemblies. The fuel channels are fabricated in the form of an array of square tubes. The walls and bottom of the inner container are then fabricated by pouring spheroidal graphite iron around the channel array.
The copper canister is fabricated either of drawn seamless tubes or by welding together two tube halves of rolled plate. A bottom is attached by an electron beam weld in such a way that the weld can be examined by ultrasonic and radiographic inspection.
After fuel has been placed in the canister, the copper shell’s lid is then attached by an electron beam weld, and leaktightness is tested by ultrasonic and radiographic inspection. The canister weighs a total of about 25 tonnes when filled with 12 BWR assemblies. A canister holds about two tonnes of fuel. It is assumed in SR 97 that approximately 8,000 tonnes of fuel will be disposed of, equivalent to around 4,000 canisters.
THMC description of processes and variables
In the complete THMC description of the canister, important variables are e.g. material composition and dimensions of copper shell and cast iron insert, which determine the strength and corrosion resistance of the canister. Important processes in the canister in the long term are copper corrosion and deformation under external loading.
2.4
Buffer/backfill
In their deposition holes, the copper canisters will be surrounded by a buffer of bentonite clay. On deposition, gaps are left for technical reasons between canister and buffer and between buffer and rock. The inner gap is filled with water and the outer with bentonite pellets and water.
After deposition, the tunnel above the deposition hole will be backfilled with a material that is adapted above all to groundwater salinity on the repository site.
Buffer material
The buffer consists of MX-80 bentonite, a natural clay from Wyoming or South Dakota in the USA. The designation MX-80 is a trade name and specifies a certain grade and particle size of dried and ground bentonite.
MX-80 bentonite consists mainly of the smectite mineral montmorillonite (65–80 percent), where the clay particles are smaller than 2 µm. Chemically, montmorillonite can be described as a polyelectrolyte, where exchangeable ions are associated with the surfaces of the negatively charged clay particles. A characteristic property of the clay is that it swells in contact with water. The exchangeable ions in MX-80 consist predomi-nantly of sodium, and the material is therefore called sodium bentonite. Water-saturated buffer contains about 25 weight-percent water. The water molecules are absorbed in the material, and water transport takes place chiefly by diffusion.
MX-80 bentonite also contains the minerals quartz (about 15 percent) and feldspar (5–8 percent). Chemically important components in addition to the minerals are carbon-ates (e.g. calcite), sulphcarbon-ates, fluorides, sulphides (e.g. pyrite), iron(II) and organic matter. After wetting, the bentonite will contain a pore water of a characteristic composition that depends on the composition of the bentonite and of the water used for wetting.
Backfill material
The backfill material consists of a mixture of bentonite clay and crushed rock. The proportions are adapted to the chemical conditions on the repository site so that the backfill will have the desired characteristics. Such site-specific adaptation has not been done in SR 97. Instead, a typical composition is used consisting of 15 weight-percent MX-80 bentonite clay and 85 weight-percent crushed rock.
THMC description of processes and variables
In the complete THMC description of buffer/backfill, important variables are e.g. material composition, dimensions, water content and density. Density in particular determines the capacity of the buffer to support the canister. Material composition and water content determine the buffer’s capacity to limit transport of dissolved corrodants in to the canister, and of any radionuclides to surrounding rock. Important buffer-related processes are water uptake, swelling and chemical transport and reaction processes.
2.5
Geosphere/site descriptions
2.5.1
Crystalline rock
The deep repository will be situated in crystalline rock of granitic composition. Granitic bedrock consists for the most part of quartz, feldspars, mica minerals and amphiboles (hornblende). In addition there are small quantities of accessory minerals, which may be of geochemical importance.
The crystalline rock is also characterized by a system of fractures. The frequency, spatial distribution, size distribution, shape and orientation of the fractures are crucial in deter-mining both hydraulic and mechanical properties in the rock. Fractures occur on all scales from microscopic fractures in the rock matrix to fracture zones, i.e. large zones of significantly elevated fracture frequency in relation to the surrounding rock. Fracture zones often constitute dominant flow paths for the groundwater, and the size of the rock movements that can occur in a fracture zone is related to the extent of the zone.
2.5.2
The three sites in SR 97
Three hypothetical repository sites are analyzed in SR 97 to illustrate actual conditions in Swedish crystalline bedrock. Data are taken from Äspö in Småland, Finnsjön in Uppland and Gideå in Ångermanland, see Figure 2-3. The sites represent three areas in stable geological settings. All three sites are relatively near the coast, and Äspö is an offshore island.
The three sites have been investigated by different experts on different occasions over a twenty-year period in studies of somewhat differing purpose and scope. The quantity of material is biggest for Äspö and smallest for Gideå.
The repositories that are analyzed on the three sites are hypothetical. None of the sites are being considered for site investigations in the ongoing siting work. To emphasize this, the names Aberg, Beberg and Ceberg will be used from now on for the sites at Äspö, Finnsjön and Gideå, respectively.
Äspö (Aberg)
At Äspö, SKB built a Hard Rock Laboratory (the Äspö HRL) between 1986 and 1995. A large quantity of data were collected before and during construction of the under-ground facility. The laboratory is used today for research purposes and data on the rock are gathered continuously from various research projects in the laboratory. Only data from the pre-investigations and from the construction phase are used in SR 97, however. Äspö is an island in the archipelago situated approximately 2 km north of the Simpevarp Nuclear Power Plant in the municipality of Oskarshamn. The landscape is flat with a thin soil layer on highlands and bogs in lowlands. Most of the area is forested, but there is also cultivated land and pastureland. The elevation difference between the highest and lowest (sea level) point in the region around Äspö is about 30 m. The rate of land uplift is 1–2 mm per year.
Figure 2-3. Data for the hypothetical repository sites Aberg, Beberg and Ceberg are taken
Approximately 1,800 million-year-old Småland granites dominate the bedrock in the region. Subordinate rock types are gabbro and diorite, as well as clasts of older rock types. A so-called younger granite, with an age of 1,400 million years, exists in the form of several massifs north and south of Äspö. There are four rock types on the actual island of Äspö: Äspö diorite, Ävrö granite, greenstone and fine-grained granite. The first two are the most common and are interpreted as variants of Småland granite.
The region contains kilometre-wide height area bordered by large valleys, which are normally interpreted as fracture zones. The frequency of regional fracture zones in and around Äspö is high compared with the surrounding region. There are few observed regional plastic shear zones. An exception is an approximately 100 m wide northeast-trending zone that runs through the central portion of Äspö.
The frequency of interpreted and observed regional and local fracture zones within Äspö is relative high.
The transmissivity of the fracture zones, as measured in boreholes on Äspö, varies between approximately 10–4 and 10–7 m2/s. The hydraulic conductivity in the rock mass
between the fracture zones on Äspö is on average around 10–8 m/s below 100 m. No
change with greater depth has been noted.
The groundwater under Äspö is non-saline down to a depth of approximately 200 metres, below which the salinity increases and is about 11,000 mg/l at a depth of 500 metres, which is twice as high as today’s Baltic Sea water.
Finnsjön (Beberg)
Finnsjön was investigated mainly during the period 1977–1978. The area was one of three that were studied to obtain data for the KBS-3 projects. Most of the knowledge that exists today from the area stems from a research project that was conducted during the years 1985–1989 in the northern part of the area. The purpose of the project was to study the geological and hydrogeological character of a flat permeable fracture zone and evaluate its importance for a repository for spent nuclear fuel.
Finnsjön is situated just north of Österbybruk, in the municipality of Tierp. The distance to the Baltic Sea is about 14 km. The landscape in and around Finnsjön is flat with rock outcrops, mires and small lakes. There are small cultivated areas in its environs. The rate of land uplift is 5–6 mm /y.
The oldest rock types in the region are approximately 1,900 million-year-old altered sedimentary and volcanic surface rock types. The latter contain iron minerals. Most of the region consists of slightly younger plutonic rocks: gabbro, diorite, tonalite, grano-diorite and granite. They have intruded into the older surface bedrock. All of the aforementioned rock types were affected by a deformation phase 1,850–1,780 million years ago. The bedrock in the Finnsjön area is dominated by granodiorite, which is normally grey with a northwesterly and steeply-dipping foliation. In tectonically affected portions the granodiorite is reddish.
The deformation phase gave rise to an extensive system of plastic regional shear zones, mainly trending northwest. This direction is also common for regional fracture zones, which, in combination with crossing zones, give rise to a block-like network of fracture zones in the region.
Finnsjön can be divided into a northern and a southern block, separated by a steeply-dipping fracture zone. The blocks are bounded on the west and east by wide regional zones. There is a gently-dipping fracture zone in the northern block with a thickness of about 100 metres whose top surface is located at a depth of 100–300 metres. The zone has not been encountered in boreholes in the southern block. Several smaller fracture zones occur within both rock blocks.
As far as the transmissivity of the fracture zones is concerned, the gently-dipping zone stands out with a transmissivity of approximately 10–3 m2/s. The equivalent value for
other fracture zones varies between approximately 10–4 and 10–8 m2/s. The hydraulic
conductivity in the rock mass between the fracture zones on levels below a depth of 100 metres averages about 10–8 m/s. Flow measurements and other observations indicate
a high flow in the upper portion of the gently-dipping zone. Calculations based on these measurements indicate a natural flow in this part of the zone of between 5 and 10 litres/ second over a cross-section of 1,000 metres. Below the upper part of the zone, similar measurements reveal a very low natural flow.
In the northern block, the groundwater is non-saline down to the top surface of the gently-dipping zone. Below that the groundwater is saline, approximately 9,000 mg/l, which is twice as much as today’s Baltic Sea water. Based on carbon-14 and tritium data, all samples from the saline water exhibit a very long residence time and a low fraction of meteoric groundwater. In the southern block, where the gently-dipping zone is missing, the groundwater is non-saline down to the end of the boreholes (about 700 m). Here modern hydrogeochemical analyses are lacking.
Gideå (Ceberg)
Gideå was investigated during the period 1981–1983. The area was one of four that was studied to obtain data for the KBS-3 projects.
The topographical relief in the region around Gideå varies from sea level to approxi-mately 300 metres, making it much greater than for Äspö and Finnsjön. The Gideå area is located approximately 100 metres above sea level and within a large highland plateau. The distance to the sea is approximately 10 km. The area is dominated by forest, bogs and mires. The rate of land uplift is about 8 mm/y.
Mineralogically, the region is dominated by sedimentary gneiss (metagreywacke) formed about 1,950–1,870 million years ago. Older and younger granite is also present in the form of large massifs in the region. The youngest rock type is sills and steeply-dipping dykes of dolerite aged 1,270–1,214 million years. Sedimentary gneiss and regimes with older granite dominate within the Gideå area. They are intersected by a system of east-westerly, steeply-dipping dolerite dykes that can be up to 15 metres wide. The direction of foliation in the gneiss varies, but is mostly east-west with a medium-steep dip. The frequency of interpreted and observed fracture zones within Gideå is relatively moderate.
The transmissivity of the fracture zones is lower than on the two other sites, varying between approximately 10–5 m2/s and 10–8 m2/s. The hydraulic conductivity in the rock
mass between the fracture zones is also relatively low, averaging around 10–10 m/s at
Gideå is situated below the highest coastline, which means that the area was covered by water after the most recent ice age. The highest parts of the area rose out of the sea about 8,000–6,000 years ago, at which time the sea was non-saline or weakly saline. Two sampled sections also contain weakly saline groundwater. Other sections have non-saline water.
2.5.3
Site-adapted repository layouts
The layout of rock caverns, tunnels and deposition positions in the repository system is based on principles first presented in the KBS-3 study. A number of possible repository locations have been proposed for Aberg, Beberg and Ceberg, and the main alternative for each site is presented below.
In Aberg, the proposed repository layout is split into two levels at depths of 500 and 600 metres so that the entire repository can be fit into the relatively limited study site, see Figure 2-4.
In Beberg, the proposed repository is located 600 m below sea level so as to avoid a dominant horizontal structure on the site with good margin, see Figure 2-5. The deposition tunnels are oriented perpendicular to the maximum horizontal stress. This direction has been chosen to avoid long intersections with water-bearing fractures with the same direction as the horizontal stress. This layout resembles the one used as an example in the SKB 91 safety assessment.
In Ceberg, the repository located at 500 m below sea level, i.e. around 600 m below the ground surface, see Figure 2-6. The deposition tunnels are oriented perpendicular to the maximum horizontal stress.
Figure 2-4. Repository layout for Aberg. The figure shows cross-sections at depths of a) 500 m
